Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Invoking flows in My Assistant and custom agents

Flow invocation in My Assistant and custom agents allows you to seamlessly access and run your Amazon Quick Flows directly within conversational AI experiences. Instead of switching between different interfaces, you can discover, run, and interact with flows while maintaining the context of your ongoing conversations.

This integration brings the power of structured workflows into natural conversation environments, making it easier to accomplish complex tasks without losing the context of your work or having to navigate between multiple applications.

How flow invocation works

Amazon Quick Flows can be invoked within two main conversational environments, each designed to provide seamless integration while maintaining the unique characteristics of the conversational experience.

Integration environments

My Assistant: The default conversational AI experience in Amazon Quick Suite that provides access to your organization's knowledge, data, and tools through natural conversation.

Custom Agents: Specialized conversational AI agents that your organization has configured for specific purposes, domains, or workflows.

Both environments allow you to access your flows while maintaining conversational context, but each has specific behaviors and capabilities tailored to their intended use.

Key benefits of flow invocation

Context preservation: Your conversation history and context are maintained when you invoke flows, allowing the flow to leverage information from your previous interactions.

Seamless transitions: Move between general conversation and structured flow running without losing your place or having to start over.

Unified experience: Access all your flows from within your regular conversational workflow, reducing the need to switch between different applications or interfaces.

Enhanced capabilities: Combine the flexibility of conversational AI with the structured power of flows to accomplish complex, multi-step tasks.

Choosing your invocation environment

The choice between My Assistant and Custom Agents depends on your specific needs and how your organization has configured these environments.

When to use My Assistant

My Assistant is ideal for general-purpose work where you need access to a broad range of organizational capabilities alongside your flows. It provides a comprehensive environment for knowledge work, data analysis, and task running.

Best for: Daily knowledge work, exploratory analysis, general task automation, and situations where you need access to multiple organizational resources and flows.

When to use Custom Agents

Custom Agents are designed for specific domains, roles, or workflows. They provide specialized conversational experiences tailored to particular use cases or organizational functions.

Best for: Domain-specific work, role-based tasks, specialized workflows, and situations where you need focused functionality within a particular area of expertise.

Getting started with flow invocation

To begin using flows within conversational agents, you'll need access to either My Assistant or Custom Agents in your Amazon Q Business environment. Your organization's administrators control access to these features and configure which flows are available in each environment.

Prerequisites

Flow access: You must have access to flows either as their creator or through sharing permissions. Only flows you can normally access will be available for invocation within agents.

Agent access: Your organization must have enabled Chat agents settings in custom permissions, and you must have appropriate permissions to use these conversational environments.

Integration enabled: Your administrators must have enabled flow invocation within the conversational environments you want to use.

Basic invocation process

Discovery: Find flows through search, recommendations, or browsing within the conversational interface.

Selection: Choose the flow you want to run explicitly.

Running: Interact with the flow conversationally or through structured interfaces, depending on your preference and the flow's design.

Completion: Return to general conversation or continue with related flows as needed.

Detailed guidance by environment

For specific information about using flows in each environment, see the detailed guides:

Each guide covers the specific behaviors, capabilities, and best practices for that environment, including getting started instructions, invocation methods, context sharing, and management features.

Common features across environments

While My Assistant and Custom Agents have unique characteristics, they share several common features for flow invocation.

Universal capabilities

Explicit invocation: Directly search for and select flows to run within the conversational interface.

Context sharing: Leverage information from your conversation history to provide inputs and context to flows.

Progress tracking: Monitor flow running progress through visual indicators and status updates.

Guardrails: Benefit from consistent security and compliance controls across both conversational and flow experiences.

History integration: Access your flow running history and conversation logs in a unified way.

Consistent user experience

Natural language interaction: Communicate with flows using natural language, just as you would in regular conversation.

Flexible running modes: Choose between conversational and structured interaction modes based on your preferences and the flow's complexity.

Error handling: Receive clear guidance when issues occur, with consistent error resolution patterns across environments.

File handling: Upload and work with files seamlessly within the conversational context when flows require file inputs.

Best practices for flow invocation

To get the most out of flow invocation in conversational agents, follow these recommended approaches.

Managing flow running

Monitor progress: Pay attention to progress indicators to understand what's happening, especially for complex or long-running flows.

Provide clear inputs: When flows request information, be specific and complete in your responses to avoid back-and-forth clarification.

Use appropriate modes: Choose between conversational and structured interaction based on the complexity of the flow and your comfort level.

Amazon Quick Suite reference